There are a couple kinds of CAES. The best ones (“advanced” methods like adiabatic and isothermal) can have 70-80% roundtrip which is pretty good – and much better than the 40-60% that “conventional” CAES is good for. Like many other things it could definitely have its place in a heterogeneous mix.

Yup, that’s why you need to keep the Bell box in the mix as basically an SFP->Ethernet converter

The only thing the Bell modem should be doing is providing your ISPs service to your new router, thus one cable from the ISP to the Bell modem, and one cable from the Bell modem to your new router. Everything else would be plugged into the new router’s LAN ports or connected via wifi to the new router. Internal traffic (from, say, your laptop to your networked printer), would only ever go through your new router. Anything that needed to go to the internet would go wired/wifi -> your new router -> the bell modem -> the internet

ok, so you have fiber-to-the-home, that SFP connector (the white cable) is a fiber optic line. So in a perfect world you have white cable -> sfp, then bell WAN port -> new router WAN port, then your new router will do its own wifi (disable the bell’s), and plug the blue ethernet into one of the new router’s LAN ports so that the new router is managing the ethernet for your home devices. That’s in an ideal world. In reality, depending on how that Bell device is set up and what kind of new router you get it might be WAN->WAN but it might also be WAN->LAN or even LAN->LAN

@[email protected] wait, do you have 2 devices from your ISP? what kind(s) are they (make/model) - they’re all a bit different - and what kind of service do you have (cable, fiber, 5g)? Normally in this config (for a cable setup, for example), you’d have:

             cable modem                          LAN Router  
| ISP | -> | [cable jack] [ethernet jack] | -> | [WAN jack] [LAN jacks]  |

If there’s a WAN port that isn’t actually used to connect that device to the ISP’s network, it’s possible it can be used to pass the traffic through (the alternative is that device might be used on more than one kind of network that does use the WAN port). But typically when you can’t use the device like a bridge, you’d have the ethernet jack from that device going to… maybe the WAN port or maybe a LAN port on your new router (depends on what it supports out-of-the-box or whether you’d be flashing it with something like OpenWRT)

The main gotcha is whether your ISP’s device can be put into “bridge mode”, which basically tells it to just act like a modem and pass the traffic along to your new router without doing anything to it. This makes everything just work, so if you self-host any services or do online gaming there won’t be issues. Otherwise, you’ll have to configure your ISP’s device to be as dumb and permissive as possible - disable DHCP so it’s not trying to give your devices IP addresses (this is your new router’s job now), and ideally put your new router into a DMZ (demilitarized zone, basically a part of your LAN that isn’t managed) so that again your new router can do all of the fancy routing and firewall stuff.

I mean… when you consider that the bulk of the world’s work products are produced by totally average people, not amazing ones, an LLM spits out completely average derivative work at an extremely high rate. Think slop speed, not slop quality (and for me, 90% of bio-slop and machine-slop are interchangeable)

Guess they should have aspired harder.

It’s funny you should mention that, because they go out of their way to specifically say that they aren’t!

Let’s just fix that headline to make it a bit more accurate:

AI backlash threatens make-believe trillion-dollar data centre boom before they have a chance to cash out, say tech leaders

At some point - hopefully soon - people will realize that LLMs simply can’t make anything new. They’re amazing at making derivative work, but that’s about it.

What kind of “real trouble” can a trillionaire even get into? Unless somebody makes a huge effort to cause him physical harm (or his family, I guess, but… yikes), he’s basically immune to consequences.

Right, though on a largeish instance with a limited blocklist you won’t see much difference (new communities mostly)

Every Lemmy instance has a built-in search tool that can be used to search across instances (make sure the slider at the top is set to “All”, it usually defaults to “Local”). Browsing All for a while will show you a decent cross-section of the most active communities (just make sure to tighten your filters). And there are some 3rd party search tools like https://lemmyverse.net/ that do periodic trawls of posts and traffic.

I’m considering thundermail, though it’s new and not perfect. Data is hosted in the EU, they don’t market your data, and it’s managed by the thunderbird subsidiary of the Mozilla group. Some basic things like E2E and at rest encryption are either half-baked or missing, but their project plan looks promising and they include calendar and contact sync as well.

The company I work for uses CodeRabbit Security and allows our team to use it for personal FOSS projects. I don’t know if it’s the best tool, but it’s the one I know (and certainly one of the more popular ones), so it’ll still do a much better job than I would manually.

Yep. honestly it doesn’t bother me if nobody else uses it (for ideological reasons or otherwise), I just figured if it’s useful to me it might be useful to others as well.

This is how I’m leaning as well. I’m no strict idealogue :)

hmm, where have I seen this before? 🤔

It’s almost like having basic appliances connected to the internet is a bad idea. Outbound sensor data? Yeah, ok, maybe. But inbound commands?